Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/stefan-jansen/claude-code-toolkit/nextgit clone --depth 1 https://github.com/stefan-jansen/claude-code-toolkitWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00008 | $0.00493 |
| Opus 5 | $0.00004 | $0.00246 |
| Sonnet 5 | $0.00002 | $0.00099 |
| Haiku 4.5 | $0.00001 | $0.00049 |
Grade A, and why
next scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
Execute Next Task
Execute pending tasks from .claude/work/ACTIVE_WORK work unit.
Arguments: $ARGUMENTS
Modes
--preview: List available tasks--status: Show progress (completed/pending/blocked)--task TASK-ID: Execute specific task--parallel N: Execute N independent tasks concurrently--parallel auto: Execute all independent tasks (max 5)- (no args): Execute next pending task
Process
-
Load State: Read
ACTIVE_WORK, verifystate.jsonexists with statusplanning_completeorimplementing -
Find Tasks: Query state.json for pending tasks with satisfied dependencies
-
Execute:
- Single task: Work directly on the task
- Parallel: Launch Task agents concurrently (single message with multiple Task tool calls)
-
Validate: Run tests, verify acceptance criteria met
-
Update State: Mark task completed in state.json, update
current_task -
Commit: Create atomic commit with task ID and description
Parallel Execution
For --parallel, find independent tasks (no unmet dependencies), launch as concurrent Task agents:
- Each agent gets: task ID, title, description, acceptance criteria
- Collect results, update state for each
- Handle partial failures gracefully (some succeed, some fail)
State File Format
{
"status": "implementing",
"current_task": "TASK-003",
"tasks": [
{"id": "TASK-001", "title": "...", "status": "completed", "dependencies": []},
{"id": "TASK-002", "title": "...", "status": "pending", "dependencies": ["TASK-001"]}
]
}
Error Handling
- No active work unit → "Run /explore first"
- No state.json → "Run /plan first"
- All tasks blocked → Show blocked reasons
- Partial parallel failure → Complete successful tasks, report failures
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 64 lines · 8 tokens per session scan A a9075f4c1a40
next is a command published in the GitHub repository stefan-jansen/claude-code-toolkit (85 stars, last pushed 1mo ago), licensed MIT. It adds 8 tokens to every session and 493 once invoked, about $0.0000 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
remember
Save something to agent memory - picks project or user scope automatically from context.
disambiguate-plan
Thorough collaborative planning with iterative disambiguation - resolves all ambiguity before planning.
pr-review
Review a GitHub pull request: analyzes changed files to select relevant reviewers, loads project and user memory, runs reviewers in parallel, lets the user pick which findings to post, then submits the review on behalf of the user via gh CLI.
dataviz
Define chart types, color encoding, and data display conventions.
ship
Pre-launch gate — six-domain checklist before any production deployment.
build
Implement the next ready task from .forge/tasks.yaml using TDD discipline.